N.-"A-'-T'-'U:..:.R=E'-V'-O::.:L::;.·..:;3.:..17;_.:..10:.._0CT:::=:..:Oc:.:B:,:E:c:Rc.clcc9.

:::85'------------REVIEW ARTICLE----------·---------=
The hepatitis B virus
Pierre Tiollais, Christine Pourcel & Anne Dejean
Unite de Recombinaison et Expression Genetique, INSERM U163, CNRS LA 271, Institut Pasteur, 28 rue du Dr Roux, 75015 Paris, France

DNA recombinant technology has radically changed hepatitis B virus (HBV) virology. The genetic
organization, transcription and replication of the virus are basically understood, structures of integrated
HBV sequences in hepatocellular carcinoma have been characterized, and new vaccines produced by
recombinant DNA technique are being developed.

HEPATITIS Bis a public health problem of worldwide import-

ance. In Far-East Asia and tropical Africa, chronic carriers of
the hepatitis B virus (HBV) represent 10% or more of the
population and chronic active hepatitis and liver cirrhosis are
major causes of mortality. Moreover, epidemiological studies
have clearly shown the importance of HBV in hepatocellular
carcinoma (HCC), one of the most common cancers in the Empty viral env elope s

/ (HBsA
gloEn(~~~~:~
a:
world 1 • In China, between 500,000 and one million new cases
appear every year. HBV in one of the few viruses known to be
•
• •
involved in a human cancer. Capsld
HBV infection is highly polymorphic, ranging from inap- · (HBcAgl
parent forms to acute hepatitis and severe chronic liver disease. DN A
The pathological consequences of the viral infection are unpre- •
dictable and the mechanism of liver damage is not fully under- The vlron
stood. The virus is not cytolytic and the host immune response Fig. 1 The hepatitis B virus particles. Left: An electron micro-
to viral antigens present at the liver cell membrane has a major graph of the viral particles present in the serum of an HBV chronic
role in the pathogenesis of HBV-related liver diseases 2 • carrier (micrograph courtesy of A. J. Zuckerman). Numerous 22-
HBV belongs to the group of animal viruses known as the nm spherical particles and filaments and one virion (or Dane
particle) are present. Right: A schematic representation of the viral
hepadna viridae 3 • The other members of this group are the particles. The virion consists of an envelope carrying HBsAg and
woodchuck hepatitis virus (WHV)4, the beechey ground squirrel a capsid carrying HBcAg and containing the circular DNA
hepatitis virus (GSHV) 5 and the Pekin duck hepatitis B virus molecule. The 22-nm spherical particles and the filaments are empty
(DHBV) 6 • All these viruses have a common structure. They are envelopes carrying HBsAg.
mainly hepatrotropic and lead to persistent virus infection. Only
HBV and WHY cause chronic active hepatitis and HCC. Analysis of its primary structure deduced from the nucleotide
sequence shows that it contains three hydrophobic sequences
Hepatitis B virus particles separated by two hydrophilic ones. Precise limits of these seg-
During HBV infection in humans, virus particles are present in ments cannot be defined, but for the three hydrophobic segments
very large quantities in the blood (Fig. 1). In some forms of they could be amino acids 7-23, 80-98 and 169-226 10 . The two
chronic infection, the serum contains only empty viral envelopes. first hydrophobic sequences probably traverse the lipid bilayer.
Occasionally, it also contains complete virions, the empty Most of the amino-acid variations observed in the different
envelopes always remaining in large excess. The presence of sequenced HBV genomes occur in the two hydrophilic domains.
complete virions in the serum is indicative of an active viral HBsAg is a conformational antigen. A dimer of two major
multiplication in the liver. The empty envelopes consist of protein molecules linked by disulphide bridges represents the
spherical or filamentous particles 22 nm in diameter. The virion structural unit that bears full HBsAg antigenicity". Reduction
of 42 nm diameter consists of an envelope and a nucleocapsid of disulphide bonds results in dissociation of the dimer and a
containing a circular DNA molecule, a DNA polymerase, a drastic decrease of HBsAg antigenicity. The segment of the
protein kinase activity and a 'DNA-linked protein'. The envelope major protein between residues 122 and 155, in the second
carries the hepatitis B surface antigen (HBsAg) and the capsid, hydrophilic domain, is an exposed 12 region and results obtained
the hepatitis B core antigen (HBcAg). When virions are present using synthetic oligopeptides have shown that this domain con-
in the blood, an additional soluble antigen related to the capsid, tains most of the HBsAg group and subtype determinants 13 - 18 .
the hepatitis B e antigen (HBeAg), is generally detected in the The middle protein is 281 amino acids long and is encoded
serum. HBsAg contains a group determinant called a and sub- by the pre-S2 region and S gene (Fig. 2) 8 • It is a glycoprotein
type determinants termed d, y, wand r. Three main serotypes, present in two forms, GP33 5 and GP365 , according to the extent
adw, adr and ayw, are commonly observed but have an unequal of glycosylation; GP33 5 contains one glycan and GP365 two
distribution worldwide 7 • They and r determinants are, respec- glycans. The 55 amino acids encoded by the pre-S2 region arc
tively, absent from Far-East Asia and from Africa. in general hydrophilic and contain a dominant epitope located
The envelope of HBV consists of proteins, carbohydrates and at the surface of the envelope 19 •20 • This epitope is disulphide
lipids. Glycoproteins are anchored in a lipid bilayer. Elec- bond-independent and apparently more immunogenic than are
trophoretic analysis and comparison of tryptic digests with the epitopes of the major protein. The pre-S2-encoded sequence
sequence of the envelope genes recently led Gerlich and his contains a receptor for polymerized human serum albumin
9
colleagues to redefine the protein composition of the envelope 8 • . (pHSA) 21 • The hepatocytes also have a receptor for pHSA and
The envelope of the virion contains three proteins called the it has been postulated that pHSA mediates the attachment of
major, the middle and the large proteins. The major protein is HBV to hepatocytes.
226 amino acids long and is encoded by the S gene (Fig. 2). It The large protein is encoded by the prc-S 1 region, pre-S2
exists in two forms, glycosylated (GP27 5 ) and non-glycosylated region and S gene (Fig. 2) and is present in a glycosylated
(P245 ). GP27 5 possesses a complex N-linked glycan at Asn 146. (GP42 5 ) and a non-glycosylated (P39s) form 9 . This protein is
© 1985 Nature Publishing Group
490
--'----------------------REVIEW ARTICLE----------"--'N""ATUR==E'-V'--'O:.-=L=--=31"-7--"10.::.....=.0CTO..::..::..::.-=B:=ER::.:....::.19=85
.c..c

of variable length according to the subtype, 389 or 400 amino

acids long for the ay and ad subtypes, respectively. GP42s
contains one N-linked glycan. At least part of the N-terminal
sequence, encoded by the pre-S 1 region, is located at the surface
of the envelope. The pre-Sl translation product may also be
involved in the attachment of HBV to hepatocytes 20 •
One virion contains 300-400 major protein molecules and
40-80 middle and large protein molecules 9 • The protein compo-
sition of filaments is identical to that of the complete virions
while the composition of the 22-nm diameter empty spherical
particles is quite different. In the serum of chronic carriers with
viral replication, the spherical particles contain both the major
and middle proteins in about the same ratio as in the virions
but with at least 20 times less of the large protein. This may be
the result of a difference between the mechanism of assembly
of virions and of empty envelopes. In the absence of viral
replication, the spherical particles contain mostly the major
protein, and < 1% of middle protein and no large protein. The
relative rates of synthesis of the three envelope proteins could
be regulated at the transcriptional level.
The nucleocapsid of HBV consists of one major protein, the
core protein (P22c), and contains a protein kinase activity cap-
able of phosphorylating the P22c protein 22 •23 • The C-terminus
of P22c is remarkable in that it is extremely rich in arginine,
serine and proline residues 24 • Presumably, the arginine-rich
sequences, in which there is a tandem repeat of the octapeptide
Ser-Pro-Arg-Arg-Arg-Arg-Ser-Gln, are involved in interactions
with the HBV genome within the nucleocapsid. Such sequences
resemble protamine and other DNA binding proteins. HBcAg
is composed of both conformational and linear determinants of
the P22c protein. HBeAg is a cryptic determinant released by
disruption of the capsid 25 , and is present in the serum as a
protein of 15,000 (15 K) relative molecular mass (M,) which
results from proteolytic cleavage of P22c. The C-terminal
sequence Thr-Thr-Val-Val was determined for serum HBeAg.
This sequence lies 33 amino acids from the end of P22c, within
the sequence Thr-Thr-Val-Val-Arg-Arg 26 • The HBV genome has
an unusual and characteristic structure (Fig. 2) 24 • It is a small
circular, partly double-stranded DNA molecule with a single-
stranded DNA region of variable length. The long or L(-) strand
is linear and of a fixed length of about 3,200 nucleotides. The
short or S( +) strand is of variable length, ranging from 50 to
100% of that of the L( - ) strand. The positions of the 5' ends
of the L( - ) and S( +) strands are fixed, while the position of
the 3' end of the S( +) strand is variable. The maintenance of Fig. 2 Structure and genetic organization of the HBV genome.
the circular structure of the genome is assured by base-pairing The broad arrows surrounding the genome represent the four large
open reading frames of the L( - ) strand transcript. The number of
of the 5' ends of the two strands. The 5' ends of the S( +) and
amino acids (aa) encoded by the coding sequences are indicated.
L(-) strands were recently mapped to positions 1,601 and 1,826
The two thin arrows surrounding the broad arrows represent the
(H. Will, personal communication). Interestingly, at both sides two major HBV mRNAs. The partial restriction map and the
of the cohesive ends there is an 11-base-pair (bp) direct repeat numbering of the nucleotides indicated on the inner circle corre-
(DR) 5'TTCACCTCTGC (Fig. 2). The two copies of this spond to the ayw 3 genome27 • DRI and DR2 are, respectively, at
sequence, which start at nucleotides 1,824 and 1,590, are termed positions 1,824 and 1,590. The positions of initiation and stop
DRl and DR2, respectively. That a similar direct repeat is found codons of the coding sequences and the 5' and 3' ends of the two
in all the hepadna viruses argues for a biological role for the major transcripts are indicated.
DR sequence (Fig. 3).
virtually all of the protein coding capacity of the virus and can
Genetic organization of HBV therefore be considered as the minus strand. The four ORFs of
The restricted host range of HBV, which infects only man and the HBV L( - ) transcript are termed S, C, P and X. The P region
chimpanzee, and the lack of a cell culture system for its propaga- overlaps with all the other three. The L( - ) strand in this pres-
tion have greatly impeded our understanding of its molecular entation is read one-and-a-half times 1°.
genetics. A general presentation of the genetic organization of The S region codes for the protein of the viral envelope and
the hepadna viridae was deduced from the comparative analysis is divided into the S gene, pre-Sl region and pre-S2 region (Fig.
of the nucleotide sequences of cloned genomes 24 • Large open 2). The S gene and pre-S2 region are of constant length in all
reading frames (ORFs) conserved among all the sequenced the HBVsubtypes. The 5' end of the pre-Sl region ofad subtypes
genomes probably represent coding regions for viral proteins contains an additional sequence of 33 or 12 bp compared with
(Fig. 2). ay subtypes 33 • The heptanucleotide sequence 5'GCTAGGG
The L(-) strand transcript of HBV contains four ORFs con- directly repeated at both extremities of the additional sequence
served in the sequences of different strains. Moreover, insertions could be related to its insertion or deletion. In general, there is
or deletions present in HBV subtypes are always multiples of about twice as much amino-acid variability in the pre-Sl and
three, allowing conservation of the reading frame 27 - 32 • In con- pre-S2 regions as in the S gene. This suggests that the amino-acid
trast, there are no conserved ORFs of any considerable length sequence of the products of the pre-S regions are less critical
in the S( +) strand transcript. The L( - ) strand thus carries for virus assembly than that of the S gene. Variations of the
© 1985 Nature Publishing Group
 ---------REVIEW AITTICLf----- ---------------4=91
_NA~T_U---'R_E_V__cOc..cLcc..-'-3'-17---'l'-'-O---'O'--'C'-'-T-'-O""BE=R-'--=-'19-"8"-5
HBV
WHY
GSHV
DHBV
Fig. 3 Genetic organization of the hepadna viruses. The broad
arrows surrounding the genomes represent the large open reading
frames of the L( - ) strand transcript. The numbers indicate the
lengths of the coding sequences. The short arrows perpendicular
to the broad arrows indicate the initiation sites of transcription of
the major viral mRN As. The asterisks indicate the common position
of the 3' end of the viral mRNAs. DRl and DR2 are respectively
indicated as 1 and 2. The sequence of the DR of the four hepadna
viruses and the distance between DRl and DR2 are represented
at bottom right.
pre-S regions could be a means of evading the host immune
system.
The C gene encodes the core protein. The N-terminal amino-
acid sequence of P22c is unknown and two AUG at positions
1,814 and 1,901 (Fig. 2) are candidates for the initiation of
translation of P22c. According to the mapping of the 5' end of
the P22c messenger RNA (H. Will, personal communication),
the AUG at position 1,901 should be considered as the initiation
codon, and we will refer to the region between AUG 1,814 and
1,901 as the pre-C region. The coding capacity of the pre-Chas
been shown in vitro using expression vectors. Initiation of trans-
lation at the pre-C initiation codon produces a hydrophobic
polypeptide bearing HBeAg and which could have a role in the
attachment of core to the viral envelope (ref. 10 and W. H.
Gerlich and W. Rutter, personal communications).
The translation product of the P region is a basic protein,
rich in histidine, of -90,000 M,, the relative molecular mass of
DNA polymerases. An amino-acid sequence comparison
between the translation product of the P region and the reverse
transcriptase ( pol g~ne product) of all retroviruses sequenced
to date and of cauliflower mosaic virus ( CaMV) reveals tracts
34
of homology in a region that overlaps the end of the S gene .
The P region, therefore, certainly encodes the viral DNA poly-
merase which possesses a reverse transcriptase activity.
The X region can code for a polypeptide of 145-154 amino
acids depending on the subtype. The function of this region is
unknown, but recently Moriarty et al. 35 have detected in HBV-
infected liver a polypeptide (P28) that reacts with antibodies
against synthetic peptides corresponding to the X region. Anti-
bodies against the translation product of this region were also
detected in sera of patients with HCC 35 •36 •
The overall structure of the genomes of the hepadna viruses
is similar (Fig. 3). The size of the WHY and GSHV genomes is
around 3,300 bp 37 •38 and that of DHBV 3,000 bp • Nucleotide
39
homologies between HBV and the three animal hepadna viruses
WHY, GSHV and DHBV are around 70, 55 and 40%, respec-
tively. WHY and GSHV sequences are very closely related, with
> 80% homology. The genetic organization of WHY and GSHV
are also very similar to each other and to that of HBV. Only
the pre-St nucleotide sequences of WHY and GSHV differ
completely from that of HBV because of the presence of large
insertions and deletions. The genetic organization of DHBV
differs somewhat from those of the mammalian hepadna viruses.
The L(-) strand transcript contains only three ORFs: S, C and
P. The 5' end of the DHBV C gene shows no homology with
the X gene of mammalian hepadna viruses. Therefore, the
DHBV C gene cannot be considered as a fusion sequence
between gene C and region X. The absence of the X region in
DHBV raises the question of its possible role in the other
hepadna viruses.
The nucleotide sequences and genetic organization of mam-
malian hepadna viruses show close homology. Nevertheless, the
pathology of GSHV, characterized by an asymptomatic per-
DR2 DR1 sistent infection and no HCC, is closer to that of DHBV than
5'TTCACCTCTGC -
+1trand
223 bp - TTCACCTCTGC to other mammalian hepadna viruses. Variations in the host
TCACCTGTGC - 212 bp - TCACCTGTGC
immune response to the different hepadna viruses may explain
these differences. Construction of recombinants between WHV
TTCACCTGTGC - 211 bp - TTCACCTGTGC
and GSHV and subsequent infection of either woodchuck or
TACACCCCTCTC - 46 bp - TACACCCCTCTC ground squirrel could help to elucidate the role of specific viral
sequences in the pathology, including HCC, of these two viruses
( C. Seeger and H. Varmus, personal communication). Moreover,
the absence of chronic hepatitis and HCC in chimpanzee infec-
ted with HBV indicates the importance of the host response in
the liver diseases.
Transcription of HBV
In the absence of cell culture systems to propagate HBV, chroni-
cally infected liver of chimpanzee constitutes the most suitable
model for studying viral transcription during virus replication.
Permanent cell lines established from human HCC containing
integrated HBV sequences and transfected cell lines with cloned
HBV DNA are also useful for studying HBV transcription in
the absence of viral replication. This last system allows the
precise control of the HBV DNA fragments and especially of
the regulatory elements introduced into the cell.
Two major HBV-specific poly(A)+ RNAs were characterized
in chimpanzee infected liver40·41 . Both are L( - ) strand tran-
scripts and are referred to as the 2.1-kilobase (kb) and the 3.5-kb
RNA (Fig. 2). The 5' end of the 2.1-kb RNA was mapped to
-20 bp upstream of the pre-S2 region and the 3' end to the very
beginning of the C gene. It covers the pre-S2 region, gene S and
region X. The 5' end of the 3.5-kb RNA was mapped to the
pre-C region (H. Will, personal communication) and the 3' end
to the same position as that of the 2.1-kb RNA. It covers the
complete genome plus an additional sequence of -100 nucleo-
tides. This implies that the 3.5-kb RNA is produced from
covalently closed double-stranded DNA and processed only at
the second run of transcription. Minor RNA species of larger
and smaller sizes were also detected. No splicing event seems
to exist in HBV transcripts, at least concerning the two major
RNAs. Cell transfection experiments with cloned HBV DNA
fragments and polymers of HBV DNA have shown clearly that
the 2.1-kb RNA is the messenger for the major protein of the
envelope and the 3.5-kb RNA the messenger for the core pro-
tein42·43. The mRNAs encoding the other viral proteins remain
unknown, but the 3.5-kb RNA probably encodes the DNA
polymerase.
Other transcripts have been characterized by in vitro experi-
ments: two polymerase II-dependent L(-)-strand transcripts,
initiated respectively at positions 1,680 and 2,810, and one
polymerase III-dependent S( +) transcript, initiated at position
1,630 and 700 bases long44•45 . The two first transcripts could be
used for translation of the pre-St and pre-C regions. Virus-
specific poly(A)+ transcripts were also characterized in hepadna
animal viruses (Fig. 3). For WHY and GSHV, two major tran-
scripts of 2.1 and 3.7 kb plus additional minor transcripts were
found in infected liver with active viral replication46· " 0. For
DHBV, a third major poly(A)+ transcript, initiated a few nucleo-
tides downstream from the initiation codon of the pre-St region,
was detected47 . Note that for both WHY and DHBV, the 2.1-kb
RNA initiates downstream from the initiation codon of the
pre-S2 region. This observation raises the question of the nature
of the transcript responsible for the translation of the pre-S2
region.
© 1985 Nature Publishing Group

~49~2--------------------REVIEW ARTICLE----------NA_T_U~R_E_V_O~L~-~3~17~l~O~O~CT~O=BE=R-'---'-'19-=-85
a u H R HBV h HBV R H u
ds- ~ 3.2 -
± pds [
- ss [
Fig. 4 State of HBV DNA in the hepatocyte. HBV-infected liver
DNA was analysed by Southern blotting using HBV DNA as a
probe. a, Free replicative forms detected in the liver from a chronic
carrier. ds, Double-stranded DNA; ±pds, partially double-
stranded DNA; -ss, minus-strand single-stranded DNA. b,
integrated HBV sequences detected in a liver tumour from a patient
with HCC. U; undigested DNA; H, HindIII digest; R, EcoRI
digest; HBV, cloned HBV DNA.
Concerning the regulatory elements which control HBV gene
transcription, Cattaneo et al. 40 have noted the presence of a
sequence resembling the late promoter of simian virus 40 (SV40)
in the pre-Sl region at position 3,122. This sequence may direct
the synthesis of the 2.1-kb RNA. No particular sequence which
could be the promoter of the 3.5-kb RNA has been noted. On
the contrary, a 'TATA'-like promoter sequence situated at an
appropriate distance upstream of the 3.7-kb RNA start site was
found in WHY and DHBV40 ·' ' 0 • The polyadenylation signal com-
mon to both mRNAs is located at position 1,916, 20 bp upstream
from the 3' end 40 •41 . Little information is available concerning
the control mechanisms of viral transcription. It is clear that the
mechanisms controlling the synthesis of the 2.1- and 3.5-kb
RNAs are different. This is based on the observation that, in
the liver during viral replication, the 2.1- and 3.4-kb RNAs are
present in about the same amount. On the other hand, in patients
with integrated viral sequences, mostly HBsAg is produced, and
in mouse cells containing integrated HBV DNA, the amount of
2.1-kb RNA is at least 10 times higher than that of the 3.5-kb
RNA42 . This suggests that the S gene promoter directs constitu-
tive expression of the S gene whereas transcription from the C
gene promoter seems to be inducible. Recently, a transcriptional
enhancer element was detected 450 bp upstream of the C gene
promoter48 . This element exhibits a preferred activity in human
hepatocytes. The tissue specificity of this regulatory element
could in part explain the low level of expression of the 3.5-kb
RNA observed in the transfected mouse cells. In addition, in
transgenic mice containing integrated HBV DNA, S gene tran-
scription takes place preferentially in the liver (C. Pourcel,
unpublished data). A tissue-specific enhancer element may also
be involved in the regulation of S gene transcription.
Multiplication cycle of HBV
HBV DNA in the hepatocyte can exist in two states, free or
integrated into the host cellular chromosome (Fig. 4) 49 •50 . Free
HBV DNA, which represents intermediate forms of replication,
is detected during the acute and some chronic stages of HBV
infection. Integrated sequences are mostly observed during
chronic virus infection and especially in HCC. Free and
integrated forms can coexist within the same sample but this
does not mean that both states coexist in the same cell.
HBV DNA has also been detected in non-hepatic tissues such
as kidney, pancreas and skin 51 •52 • Extra-hepatic free viral
sequences were also found in the pancreas and kidney of infected
© 1985 Nature Publishing Group
~o
'
r
•
J Supercoiled DNA t
•
Tran~ ',
•
e P ~ 2.1 kb RNA
'
C P ~ ' ' 3.5kbRNA
~ ~
•
Fig. S Replication cycle. (I) Entry into the cell of an extracellular
virion containing the HBV genome with a short capped RNA {cp),
the DNA polymerase (e) and the DNA-linked protein (*). DR!
and DR2 are indicated by I and 2. (2) Complete double-stranded
open circular DNA. (3) Supercoiled covalently closed circular
DNA. (4) Pregenome packaged into a nucleocapsid. (5) core
containing the complete (-)strand DNA. (6) Synthesis of(+)
strand. (7) Coated nucleocapsid giving rise to extracellular virion.
Pekin ducks 53 • The presence of free and integrated forms of
HBV DNA in bone marrow and mononuclear blood cells during
both the acute and chronic stages of HBV infection raises
questions about the role of these cells in the pathogenesis of
hepatitis B54 •55 • It has been suggested that infection of mononu-
clear cells would modify the immune response 56 •
The replication mechanism of hepatitis B viruses, discovered
by Summers and Mason for DHBV 57 •58 and confirmed later for
HBV 59 - 62 , differs strikingly from that of other DNA viruses. The
central feature is the use of a RNA copy of the genome as
intermediate in replication. The replication cycle involves a
reverse transcription step resembling that of retroviruses.
In Pekin ducks 6 hours after infection, DHBV DNA is found
in the liver, in both supercoiled and relaxed circular forms.
Slightly later, RNA is detected, followed by the appearance of
single-stranded DNA 58 .
The HBV replication cycle is believed to consist of the follow-
ing steps. After penetration of the virus into the hepatocyte, the
DNA reaches the nucleus and is converted to complete open
circular double-stranded DNA and then to supercoiled DNA
(Fig. 5). The minus strand is transcribed by the cellular RNA
polymerase to the 3.5-kb RNA previously described and called
the pre-genome. Multiple copies of the pre-genome are synthe-
sized from each genome. The pre-genome is encapsidated with
the DNA polymerase and the DNA-linked protein. The minus-
strand DNA is then synthesized by reverse transcription of the
pre-genome. Initiation of reverse transcription takes place at
DRl and the DNA-linked protein covalently bound to the 5'
end of the minus strand probably acts as a primer for minus-
strand synthesis 63 . During DNA synthesis, the pre-genome is
degraded by an RNase H-like activity and the product of reverse
transcription is a full-length minus DNA strand. A small RNA
fragment of about 20 bp, probably derived from the 5' end of
the pre-genome, is then used to prime the synthesis of the plus
strand at DR2 (J.M. Lien and W. Mason, personal communica-
tion). Complete synthesis of this strand is not necessary for
coating of the nucleocapsids and export of virus particles. This
explains why infectious particles contain DNA with a single-
stranded region.
The replication cycle of hepadna viridae in some ways
resembles that of retroviruses and of CaMV. All these viruses
involve a reverse transcription step, but whereas the virion of
retroviruses contains RNA and the intermediate form of replica-
tion is integrated DNA, the virion ofhepadna viruses and CaMV
contains DNA and the intermediate replication form is RNA.
NATURE VOL. 317 10 OCTOBER 1985
REVIEWARTICLE-------- ----
_a.
Recruit No. of
status patients (No of cssesl
Fig. 6 Epidemiology of hcpatocel- studied
lular carcinoma'. a, Relative risk of
HCC in relation to recruitment
HBsAg status. b, Cause of death in HBsAg+ 3,454
relation to recruitment HBsAg status.
19,253
• During 6.2 years follow-up .
Integrated sequences have never been detected during DHBV
replication (C. E. Roiger and J. Summers, personal communica-
tion). An integration step of the genome is therefore probably
not necessary for hepadna virus replication. Minus-strand syn-
thesis of retrovirus and CaMY DNA is primed by a transfer
RNA while hepadna DNA minus-strand synthesis is probably
primed by a protein. The order of the three retroviral genes,
gag, pol, env, can be compared to that of the order of the hepadna
virus genes C, P and S, but whereas retrovirus proteins are
translated from mRNA produced by splicing from a single
primary transcript, hepadna proteins are translated from at least
two unspliced primary transcripts. For all three classes of virus,
promoters are closely followed by polyadenylation signals that
are passed over during the first transcription run.
HBV and hepatocellular carcinoma
Evidence that HBY causes HCC comes from epidemiological
studies. There is a striking worldwide geographical correlation
between the prevalence of HBsAg chronic carriers and the
annual incidence of HCC 64 • Moreover, retrospective epidemio-
logical studies have shown that HBsAg is more frequently
observed in the serum of patients with HCC than in the control
population 64 • The strongest evidence for the role of HBV in
HCC comes from prospective epidemiological studies perfor-
med in Taiwan by Beasley et al.65 , who studied male subjects
aged between 40 and 59 y. A male population was chosen
because HCC is about three times more frequent in males than
females . The results showed that the relative risk to HBsAg
carriers of developing HCC was 217 compared with non-carriers
(Fig. 6). Furthermore, 51 % of deaths among HBsAg carriers
were caused by cirrhosis or HCC compared with only 2% among
the control population. Among the non-carriers, 90% possessed
an HBY marker, anti-HBs and/ or anti-HBc antibodies. The high
incidence of HCC is therefore clearly related to the HBsAg
chronic carrier state and not to HBY infection per se.
We can conclude that although the data do not preclude a
role for other factors (such as aflatoxin) in the development of
HCC, additional factors other than HBV chronic infection are
not necessary to explain the epidemiological studies. It is also
of interest that HCC developed generally after a history of
cirrhosis over a long period (30-40 y) of chronic carriage.
However, additional evidence linking HBY and HCC comes
from the WHV/ woodchuck model. It has been possible to induce
HCC experimentally in woodchuck by inoculation with WHY
at birth (J. Gerin, personal communication). This result provides
experimental support for an oncogenic role of WHV and vali-
dates the use of WHY as a pathogenic model for HBV.
The structure of the HBY DNA sequences present in HBsAg-
positive HCC has been studied by Southern blotting. Both
human HCC-derived cell lines and human tumour
samples 49 •50 •66 - 72 have been analysed. In both cases, HBV DNA
sequences integrated into the host genome were nearly always
detected. For any given tumour sample, the number of integra-
tion sites varied from 1 to 12. The presence of only free viral
DNA without integrated sequence, as observed in the chronic
carrier stage, has not been reported for HCC. However, free
HBY DNA with intermediate forms of replication in addition
to the integrated sequences are occasionally observed in the
early stage of tumour growth. Virus replication, as well as virus
© 1985 Nature Publishing Group
b
Q
HCC HCCIIO0.000• Relative HBsAg carriers Non-carriers
risk
HCC
-40%
113 3,272 217
HCC <1%
3 16 Hepatic '-Cirrhosis Cirrhosis 1.5%
necrolls 15%
gene expression, are generally absent in the poorly differentiated
hepatocytes of advanced tumours 7 3 . It is not known why HBY
replication and expression should be absent from HCC cells.
Two hypotheses can be envisaged; either cells harbouring HBY
yet without replication/ expression are selected during tumour
growth or the replication/ expression decreases gradually in
proportion to the dedifferentiation of the cell. HBcAg has been
suggested as the target for the cytotoxic immune response 74. If
so, hepatocytes unable to express HBcAg would escape the
immune response and be selected during carcinogenesis.
To investigate further the role of HBV in HCC, the presence
of HBV DNA in HBsAg-negative HCC was determined. In a
significant number of cases, integrated HBV DNA sequences
were detected even in patients without any HBV serological
marker75 - 78 • These data offer new evidence for an oncogenic
role of HBV and indicate that integration may play a part in
tumour formation; however, the exact significance of integrated
HBV sequences is unclear. Indeed, HBV sequences are not
indicative of tumorous tissue. HBY DNA was observed both in
non-tumorous parts of the livers with HCC and in chronic
hepatitis or cirrhosis without HCC 50 •79 • In addition, when the
tumorous and the non-tumorous parts of liver were analysed
simultaneously, the integration patterns were generally different
but in a very few cases the tumorous and non-tumorous tissues
shared the same pattern.
The structure of integrated HBV DNA sequences present in
tumour samples 80 •81 and in HCC-derived cell lines 82 - 85 was
determined by DNA cloning. Viral sequences can be complete
genomes or subgenomic fragments that are often rearranged.
Such complex rearrangements have already been described for
integrated viral sequences of SV40 and adenovirus 86 •87 • The
mechanism of HBY integration remains unclear; however,
Dejean et al. 81 have reported the existence of a virus-specific
site of integration. In two independent liver tumours, one of the
two host-virus DNA junctions mapped within either DRl or
DR2. The HBV DNA structure which precedes the integration
event is unknown. If the substrate of integration is linear or an
open circle, integration through the DR sequence could suggest
integration at the ends of the HBV DNA strands. If the substrate
of integration is a covalently closed circular form, the result
observed implies a specific recombination event within the DR.
C. Shih (personal communication) reported similar results and
found in, two hepatoma samples, virus-cell and virus-virus junc-
tions within a few nucleotides of a DR sequence. Taken together,
these data show that HBV DNA can integrate via a specific viral
DNA sequence. It is nevertheless possible that a nonspecific
mechanism of integration could occur in addition to that involv-
ing DRl and DR2. No evidence for a specific cellular integration
site or for extensive sequence homology between viral and
cellular DNA at this site has been reported to date.
The role of HBV in HCC is not understood at the molecular
level. HBV may act as an initiator by inducing continuous liver
necrosis and regeneration. Integration could occur at random
and cells capable of escaping the immune response would
become targets for further events of carcinogenesis. By contrast,
HBV integration may have a direct role in liver cancer. However,
none of the classical mechanisms of viral oncogenesis has so
far been demonstrated. The existence of a viral transforming
gene would imply the existence of a conserved viral sequence
_4'4
____________________ REVIEW ARTICL.£---------'--NA_nJc.......c.R=E_V~O=L.=·3'-"1_7-'-'10'-0CT-'--'-' -"O=B=ER-'-'-'-'19=8j
in the different tumorous cells. No peculiar region of the HBV
genome was conserved in the different clones of integrated HBV
sequences in human HCCs 81 - 84 and integrated WHY sequences
in woodchuck HCCs 88 • In addition, transformation experiments
with HBV DNA have been unsuccessful. It is also possible that
a hybrid or partial gene product could induce transformation.
Recently, Freytag von Loringhoven et al. 89 characterized hybrid
HBV-host transcripts in the hepatoma-derived cell line
PLC/PRF5. The slow transforming retroviruses, which do not
contain a viral oncogene, act as insertional mutagens. The
integration of the provirus can activate the expression of a
neighbouring cellular oncogene or disrupt a cellular regulator
gene, preventing its expression. The two models require that the
viral integration occur at approximately the same site in the
host genome. Thus far, there is evidence neither for HBV integra-
tion in a distinct chromosomal region nor for integration adja-
cent to any of the known cellular oncogenes. Further characteriz-
ation of integrated HBV sequences and their flanking sequences
in HCC might allow us to find a common element between the
different integration sites and hence clarify the process of HBV
oncogenesis.
Recombinant DNA and a future vaccine
The worldwide distribution of HBV infection with the con-
sequence of chronic liver disease and HCC implies the need for
an efficient and safe vaccine. Compared with more conventional
antiviral vaccines, the hepatitis B vaccine is unique in that HBV 100.000 h C
cannot be propagated in tissue culture. The vaccine is obtained 50,000
from plasma of healthy HBV chronic carriers. It consists of
purified defective 22-nm HBsAg particles formulated in an 8::; 10,000
alume adjuvant. Vaccination strategies differ according to the j 5,000
level of endemicity of the affected population. In countries with
low endemicity such as in Western Europe or North America, g
vaccination is restricted to individuals with an established high .
CD
risk of acquiring hepatitis B. These populations are mainly health I

personnel, dialysis patients, drug addicts, male homosexuals c' 100

and infants of HBsAg-positive mothers. On the other hand, in
developing countries with high endemicity such as in Far-East
Asia and tropical Africa, a programme of universal vaccination
at infancy and early childhood is desirable, if not currently
feasible.
The effectiveness and safety of the present hepatitis B vaccine
have been demonstrated both in high-risk adult populations 90 - 92
and in newborn infants 93 - 95 • Nevertheless, alternative methods
for vaccine production are needed. This is because of limitations
in the availability of human serum, the need for elaborate
purification and inactivation procedures and the requirement
of an inocuity test in chimpanzees. Plasma-derived vaccines are
very expensive and therefore beyond the financial possibilities
of many countries which need mass infant vaccination pro-
grammes.
Skelly et al. 96 have prepared polypeptides in water-soluble
mice Ile form from HBsAg particles. This subunit vaccine induces
antibodies at a higher titre than HBsAg particles but still relies
on the availability of human sera. For this reason, the use of
synthP.tic oligopeptides carrying HBsAg determinants is an
attractive alternative. Unfortunately, the low value of the
immune response so far obtained 13 - 18 limits the use of this
technique. Within this context, development of recombinant
DNA vaccines is of great importance. Several approaches can
be used and the choice of the technique depends on the knowl-
edge of the structure and the immunogenicity of the antigenic
determinants present on the surface of the envelope, mainly
HBsAg but also pre-S-encoded epitopes. HBsAg is a conforma-
tional antigen and full immunogenicity of HBsAg is highly
dependent on its tertiary structure 11 • Unlike HBcAg 97 , HBsAg
is not assembled as particles in bacteria and only bacterium-
HBV hybrid polypeptides showing HBsAg antigenicity were
obtained 98 - 100 • A eukaryotic system such as mammalian cell 101
or yeast 102 capable of correctly synthesizing native HBsAg
particles is therefore required.
At present, the most advanced recombinant vaccine is a vac-
© 1985 Nature Publishing Group
< 50
10 24 2M 10 2• 2no
Days pos1 - ,mmun1za1,on
Fig. 7 Immunogenicity of HBsAg 22-nm particles containing the
pre-S2 region translation product 107 • a, Electron micrograph of
HBsAg 22-nm particles synthesized in CHO cells. b, Production
of antibodies to HBsAg (S) and the pre-S2 region translation
products (pre-S) in a HBsAg responder mouse strain (strain BIO).
c, Production of antibodies in a HBsAg non-responder mouse
strain (strain BIOS). Dashed curves, anti-HBs response of mice
immunized with HBsAg particles lacking the pre-S2 region deter-
minants. Mouse immunizations were at days O and 24. The term
'2nd' on the x axis indicates the secondary immune response.
cine produced in yeast. Valenzuela et al. 102 , using an expression
vector containing the S gene placed under the control of the
yeast alcohol dehydrogenase I promoter, obtained the synthesis
of HBsAg particles in Saccharomyces cerevisiae. Miyanohara et
a/. 103 obtained a comparable result using the acid phosphatase
promoter. The particles had the same density as particles from
human serum but were heterogeneous in size; they contained
the major protein only in its non-glycosylated form. Injected
into chimpanzees, ttiese HBsAg particles induced anti-HBs anti-
bodies that confered immunity to HBV infection 104 • Human
trials are now in progress.
Production of HBsAg 22-nm particles in mammalian cells is
a second alternative; the particles obtained are glycosylated and
secreted in the cell supernatant1° 1 . An increase in HBsAg pro-
duction was obtained by the following method 105 • 106 . A plasmid
containing the cDNA encoding dihydrofolate reductase (dhfr)
and HBV sequences was introduced into Chinese hamster ovary
(CHO) cells lacking dhfr. Selection of dhfr-positive clones resis-
tant to methotrexate resulted in gene amplification of both dhfr
cDNA and HBV sequences, and an increase of HBV DNA
transcription. Using this procedure, Michel et al. 105 have con-
structed a plasmid containing the S gene and the pre-S2 region
_NA_TU_R_E_·V_O_L_._31_7_1_0_0Cf_O_B_E_R_19_8S_ _ _ _ _ _ _ _ _ _ REVIEW ARTIC
placed under the control of both the SV40 early promoter and
the promoter localized in the pre-S 1 region. The HBsAg 22-nm
particles obtained (Fig. 7) contained both the major protein in
its two forms, glycosylated and non-glycosylated, and the gly-
cosylated middle protein in a ratio comparable to that of virion.
The particles, homogeneous in size, contained both the HBsAg
determinants and the pHSA receptor activity, as a result of the
presence of the pre-S2 region translation product. Studying the
murine immune response to these particles Milich et al. showed
that the pre-S2 region product is significantly more immunogenic
than HBsAg and that immunization of a HBsAg non-responder
strain with these particles can circumvent non-responsiveness
to HBsAg (Fig. 7) 107 . Because the appearance of anti-pHSA
receptor antibodies seems to be a highly reliable criterion for
viral clearance, the HBsAg particles obtained in CHO cells may
constitute a particularly efficient vaccine.
The third approach for a recombinant vaccine is to construct
a live vaccine, using as vector a virus capable of replicating in
humans. Smith et al. have constructed a live vaccine using a
vaccinia virus as vector 108 • The S gene was placed under the
control of the vaccinia virus early promoter. Cells infected with
the vaccinia virus recombinants synthesized and excreted
I. Beasley. R. P. & Hwang, L. Y. in Viral Hepatitis and Liver Disease (ed. Vyas, G. N.)
209-224 (Grune & Stratton, New York, 1984).
2. Chisari, F. V. (ed.) Advances in Hepatitis Research (Masson, New York, 1984).
3. Robinson. W. S., Marion, P., Feitelson, M. & Siddiqui, A. in Viral Hepatitis 1981 int. Symp.
(eds Szmuness, W., Alter, H. J. & Maynard, J. E.) 57-68 (Franklin Institute Press,
Philadelphia, 1982).
4. Summers, J., Smolec, J. & Snyder. R. Proc. natn. Acad. Sci. U.S.A. 75, 4533-4537 (1978).
5. Marion, P. L., Oshiro. L. S., Regnery, D. C., Scullard, G. H. & Robinson, W. S. Proc.
natn. Acad. Sci. U.S.A. 77, 2941-2945 (1980).
6. Mason, W. S., Seal, G. & Summers, J. J. Virol. 35, 829-836 (1980).
7. Courouce-Pauty, A. M., Plancon, A. & Soulier, J.P. Vox Sang. +4, 197-211 (1983).
8. Stibbe, W. & Gerlich, W. H. J. Virol. 46, 626-628 (1983 ).
9. Hccrmann, K. H. et al. J. ViroL 52, 396-402 (1984).
10. Tiollais, P., Wain-Hobson, S. in Advances in Hepatitis Research (ed. Chisari, F. V.) 9-20
(Masson, New York, 1984).
11. Mishiro, S. et al. J. Immun. 124, 1589-1593 (1980).
12. Peterson, D. L., Nath, N. & Gavilanes, F. J. biol. Chem. 257, 10414-10420 (1982).
13. Lerner, R. A. et al. Proc. natn. Acad. Sci. U.S.A. 78, 3403-3407 (1981).
14. Dreesman, G. R. et al. Nature 295, 158-160 (1982).
15. Prince, A. M., lkram, H. & Hopp, T. P. Proc. natn. Acad. Sci. U.S.A. 79, 579-582 (1982).
16. Bhatnagar, P. K. et al. Proc. natn. Acad. Sci. U.S.A. 79, 4400-4404 (1982).
17. Neurath, A. R., Kent, S. B. H.& Strick, N. Proc. natn. Acad. Sci. U.S.A. 79, 7871-7875 (1982).
18. Gerin, J. L. et al. Proc. natn. Acad. Sci. U.S.A. 80, 2365-2369 (1983).
19. Neurath, A. R., Kent, S. B. H. & Strick, N. Science 224, 392-394 (1984).
20. Neurath, A. R., Kent, S. B. H., Strick, N., Taylor, P. & Stevens, C. E. Nature 315, 154-156
(1985).
21. Machida, A. et al. Gastroenterology 86, 910-918 (1984).
22. Albin, C. & Robinson, W. S. J. Virol. 34, 297-302 (1980).
23. Petit, M.A. & Pillot. J. J. Virol. 53, 543-551 (1985).
24. Tiollais, P., Chamay, P. & Vyas, G. N. Science 213, 406-411 ( 1981 ).
25. Ohori, H., Shimizu, N., Yamada, E., Onodera. S. & Ishida, N. J. gen. Virol. 65, 405-414
(1984).
26. Takahashi, K. et al. J. lmmun. 130, 2903-2907 (1983).
27. Galibcrt, F.• Mandart, E., Fitoussi, F .• Tiollais, P. '& Charnay, P. Nature 281, 646-650
(1979).
28. Pasek, M. et al. Nature 282, 575-579 (1979).
29. Valenzuela, P., Quiroga, M., Zaldivar, J., Gray, P. & Rutter, W. J, in Animal Viral Genetics
(eds Fields, B., Jaenisch, R. & Fox, C. F.) 57-70 (Academic, New York, 1980).
30. Ono Y. et al. Nucleic Acids Res. II, 1747-1757 (1983).
31. Fujiyama, A. et aL Nucleic Acids Res. II, 4601-4610 (1983).
32. Kobayashi, M. & Koike, K. Gene 30, 227-232 (1984).
33. Tiollais, P. et al. in Viral Hepatitis and Liver Disease (ed. Vyas, G. N.) 49-65 (Grune &
Stratton, New York. 1984).
34. Toh, H., Hyashida, H. & Miyata, T. Nature 305, 827-829 (1983).
35. Moriarty, A. M., Alexander, H. & Lerner, R. A. Science 227, 429-433 (1985).
36. Kay, A., Mandart, E., Trepo. C. & Galibert, G. EMBO J. 4, 1287-1292 (1985).
37. Galibert, F., Chen, T. N. & Mandart, E. J. Virol. 41, 51-65 (1982).
38. Seeger, C., Ganem, D. & Varmus, H. E. J. Virol. 51, 367-375 (1984).
39. Mandart, E., Kay, A. & Galibert, F. J. Virol. 49, 782-792 (1984).
40. Cattaneo, R., Will, H., Hernandez, N. & Schaller, H. Nature 305, 336-338 (1983).
41. Cattaneo, R., Will, H. & Schaller, H. EMBO J. 3, 2191-2196 (1984).
42. Pourcel, C. et al. J. Virol. 42, 100-105 (1982).
43. Gough, N. J. molec. Biol. 165, 683-699 (1983).
44. Rall, L.B., Standring, D. N., Laub, 0. & Rutter, W. J. Molec. cell. Biol. 3, 1766-1773 (1983).
45. Standring, D. N., Rall, L.B., Laub, 0. & Rutter, W. J. Molec. cell. Biol. 3, 1774-1782 (1983).
46. Moroy, T., Etiemble, J., Trepo, C., Tiollais, P. & Buendia, M.A. EMBO J. 4, 1507-1514
(1985).
47. Buscher, M., Reiser, W., Will, H. & Schaller, H. Cell 40, 717-724 (1985).
48. Shaul, Y., Rutter, W. J. & Laub, 0. EMBO J. 4, 427-430 (1985).
49. Brechot, C. et al. Hepato/ogy 2, 26S-34S (1982).
50. Shafritz, D. Hepatology 2, 35S-41S (1982).
51. Dejean, A., Lugassy. C., Zafrani, S., Tiollais, P. & Brechot, C. J. gen. Virol. 65, 651-655
(1984).
52. Siddiqui, A. Proc. natn. Acid. Sci. U.S.A. 80, 4861-4864 (1983).
53. Halpern, M. S. et al. Proc. natn. Acad. Sci. U.S.A. 80, 4865-4869 (1983).
54. Romet-Lemonne, J. L. et aL Science 221, 667-669 (1983).
55. Pontisso, P., Poon, M. C., Tiollais, P. & Brechot, C. B. med. J. 288, 1563-1566 (1984).
56. Elfassi, E. et al. Proc. ·natn. Acad. Sci. U.S.A. 81, 3526-3528 (1984).
57. Summers, J. & Mason, W. S. Cell 29, 403-415 (1982).
© 1985 Nature Publishing Group
49S
HBsAg particles; infected rabbits produced antibodies to
HBsAg. Chimpanzees infected with the recombinant vaccine
did not develop an anti-HBs response but were protected against
hepatitis B following challenge with HBV109 •
Which of these three hepatitis B recombinant vaccines will
constitute for the near future the most appropriate alternative
to the present hepatitis B vaccine? This question is not easily
answered because the final choice of vaccine depends on several
parameters: the immunogenicity of the particles, the possibilities
of an industrial transfer of the technical procedure and the safety
of the final product. In any case, twenty years after the funda-
mental discovery of HBsPg by B. S. Blumberg 111 , the future
recombinant DNA vaccine against hepatitis B will be the first
example of the application of genetic engineering to human
vaccination as well as the first example of a vaccine against an
infectious disease related to cancer.
The nomenclature used for HBV proteins and genes is that
agreed upon a Cold Spring Harbor meeting (May 1985).
We thank Kallee Gehring, Simon Wain-Hobson and Christian
Brech->t for revising this manuscript. This work was supported
by NIH grant CA97300-02, the Economic European Community
and the Association pour la Recherche contre le Cancer.
58. Mason, W. S., Aldrich, C., Summers, J. & Taylor, J. M. Proc. natn. Acad. Sci. U.S.A. 79,
3997-4001 (1982).
59. Blum, H. E., Haase, A. T., Harris,_ J. D., Walker, D. & Vyas, G. N. Virology 139, 87-96
(1984).
60. Fowler, M. J. F., Monjardino, J., Tsiquaye, K. N., Zuckerman, A. J. & Thomas, H. C. J.
med. Viral. 13, 83-91 (1984).
61. Miller, R.H., Tran, C. T. & Robinson, W. S. Virology 139, 53-63 (1984).
62. Miller, R.H., Marion, P. L. & Robinson, W. S. Virology 139, 64-72 (1984).
63. Molnar-Kimber, K. L., Summers, J. W. & Mason, W. S. J. Virol. 51, 181-191 (1984).
64. Szmuness, W. Prog. med. ViroL 24, 40-69 (1978).
65. Beasley, R. P., Hwang, L. Y., Lin, C. C. & Chien, C. S. Lancet vol. 2 1129-1133 (1981).
66. Marion, P. L., Salazar, F. H., Alexander, J. J. & Robinson, W. S. J. Virol. 33, 795-806 (1980).
67. Chakraborty, P. R., Ruiz-Opazo, N., Shouval, D. & Shafritz, D. A. Nature 286, 531-533
(1980).
68. Brechot, C., Pourccl, C., Louise, A., Rain, B. & Tiollais, P. Nature 286,.533-535 (1980).
69. Edman, J. C., Gray, P., Valenzuela, P., Rall, L. B. & Rutter, W. J. Nature 286, 535-538
(1980).
70. Chen, D.S., Hoyer, B. H., Nelson, J., Purcell, R.H. & Gerin, J. L. Hepatology 2, 42S-45S
(1982).
71. Shafritz, D. A., Shouval, D., Sherman, H. I., Hadziyannis, S. J. & Kew, M. C. New Engl.
J. Med. 305, 1067-1073 (1981).
72. Hino, 0. et al. Hepatology 4, 90-95 (1984).
73. Hirohashi, S., Shimosato, Y., lno, Y. & Kishi, K. J. natn. Cancer Inst. 69, 565-568 (1982).
74. Eddelston, A. L. W. F., Mondelli, M., Mieli- Vergani, G. & Williams, R. Hepatology 2,
112S-l27S (1982).
75. Brechot, C. et aL New Engl. J. Med. 306, 1384-1387 (1982).
16. Figus, A. et al. Heparo/ogy 4, 364-368 (1984).
77. Hino, 0., Kitagawa, T. & Sugano, H. Int. J. Cancer 35, 5-10 (1985).
78. Brechot, C. et aL New EngL J. Med. 312, 270-276 (1985).
79. Brechot, C. et aL Lancet U, 765-768 (1981).
80. Dejean, A., Brechot, C., Tiollais, P. & Wain-Hobson, S. Proc. natn. Acad. Sci. U.S.A. 80,
2505-2509 (1983 ).
81. Dejean, A., Sonigo, P., Wain-Hobson, S. & Tiollais, P. Proc. natn. Acad. Sci. U.S.A. 81,
5350-5354 (1984).
82. Koshy, R. et al. Cell 34, 215-223 (1983).
83. Shaul, Y .. et al. J. Virol. 51, 776-787 (1984).
84. Mizusawa, H. et al. Proc. natn. Acad. Sci. U.S.A. 82, 208-212 (1985).
85. Ziemer, M., Garcia, P., Shaul, Y. & Rutter, W. J. J. Virol. 53, 885-892 (1985).
86. Botchan, M., Topp, W. & Sambrook, J. Cell 9, 269-287 (1976).
87. Sambrook, J. et al. Cold Spring Harb. Symp. quant. Biol. 4-4, 569-584 (1980).
88. Ogston, C. W. Jonak, G. J., Rogler, C. E., Astrin, S. M. & Summers, J. Cell 29, 385-394
(1982).
89. Freytag von Loringhoven, A., Koch, S., Hofschneider, P. H. & Koshy, R. EMBO J. 4,
5
QO_ Cr~~~j;f /1l,~;rl"nn•1 i. 4S~-4~9 ( 19XI 1.
91. Laplanche, A. et aL Lancet i, 222 (1982).
92. Szmuness, W. et al. New Engl. J. Med. 307, 1481-1486 (1982).
93. Maupas, P. ti al. Lancet I, 289-292 (1981).
94. Lee, G. C. Y. ti al. J. infecr. Dis. 148, 526-529 (1983).
95. Prozesky, 0. W. et al. J. infect. Dis. 1, Suppl. I, 53-57 (1983).
96. Skelly, J., Howard, C.R. & Zuckerman, A. J. Nature 290, 51-54 (1981).
97. Cohen, B. J. & Richemond, J. E., Nature 296, 677-678 (1982).
98. Charnay, P., Gervais, M., Louise, A., Galibert, F. & Tiollais, P. Nature 286, 893-895 (1980).
99. Edman, J.C., Hallewell, R. A., Valenzuela, P., Goodman, H. M. & Rutter, W. J. Nature
291, 503-506 (1981).
100. Fujisawa, Y. er al. Nucleic Acids Res. II, 3581-3590 (1983).
IOI. Dubois, M. F. et al. Proc. natn. Acad. Sci. U.S.A. 77, 4549-4553 (1980).
102. Valenzuela, P., Medina, A., Rutter, W. J., Ammercr, G. & Hall, B. D. Nature 298, 347-350
(1982).
103, Miyanohara, A. Proc. natn. Acad. Sci. U.S.A. 80, 1-5 (1983).
104. McAlccr, W. J. et al. Nature 307, 178-180 (1984).
105. Michel, M. L. et al. Proc. natn. Acad. Sci. U.S.A. 81, 7708-7712 (1984).
106. Patzer, E. J. et al. in Viral Hepatitis and Liver Disease (ed. Vyas, G. N.) 477-485 (Grune
& Stratton, New York, 1984).
I07. Milich, D. R. et al. Sci,nce 228, 1195-1199 (1985).
108. Smith, G. L., Mackel!, M. & Moss, B. Nature 302, 490-495 (1983).
109. Moss, B., Smith, G. L., Gerin, J. L. & Purcell, R. H. Nature 311, 67-69 (1984).
I 10. Enders, G. H., Ganem, D. & Varmus, H. C,11 (in the press).
Ill. Blumberg, B. S., Alter, H.J. & Visnich, S. J. Am. med. Ass. 191, 541-546 (1965).